Cascading is a distribution technique in which data passes through one or more intermediate systems before reaching its destination. MIMIX supports cascading in both its user journal and system journal replication paths. However, the paths differ in their implementation.
Data can pass through one intermediate system within a MIMIX installation. Additional MIMIX installations will allow you to support cascading in scenarios that require data to flow though two or more intermediate systems before reaching its destination. Figure 18 shows the basic cascading configuration that is possible within one MIMIX installation.
Figure 18. Example of a simple cascading scenario
To enable cascading you must have the following:
-
Within a MIMIX installation, the management system must be the intermediate system.
-
Configure a data group between the originating system (a network system) to the intermediate (management) system. Configure another data group for the flow from the intermediate (management) system to the destination system.
-
In environments that implement cascading distributions and involve referential constraints, if any data group in the cascade chain is part of a resource group that uses multithreaded database apply processing, all subsequent downstream data groups must also be in multithreaded resource groups.
-
Specify *NO for the Lock on apply parameter of a data group definition whose target is the intermediate system. This is especially important for *SYSJRN data groups or for files that are being replicated by object only.
-
For user journal replication, you also need the following:
-
The data groups should be configured to send journal entries that are generated by MIMIX. To do this, specify *SEND for the Generated by MIMIX activity element of the DB journal entry processing (DBJRNPRC) parameter. When this is the case, MIMIX performs the database updates.
-
If it is possible for the data to be routed back to the originating or any intermediate systems, you need to use keyed replication. Keyed replication is not supported in data groups that use multithreaded database apply processing.
Cascading may be used with other data management techniques to accomplish a specific goal. Figure 19 shows an example where the Chicago system is a management system in a MIMIX installation that collects data from the network systems and broadcasts the updates to the other participating systems. The network systems send unfiltered data to the management system. Figure 19 is a cascading scenario because changes that originate on the Hong Kong system pass through an intermediate system (Chicago) before being distributed to the Mexico City system and other network systems in the MIMIX installation. Exit programs are required for the data groups acting between the management system and the destination systems and need to prevent updates from flowing back to their system of origin.
Figure 19. Bi-directional example that implements cascading for file distribution.